TRANSCRIPTEnglish

wtf wtf wtf

21m 36s3,104 words477 segmentsEnglish

FULL TRANSCRIPT

0:00

Well, the market's dumping. Why is the

0:03

market dumping? We got the answer for

0:06

you. It's because of this news out of

0:09

Oracle, and it's absolutely

0:12

bad. It is the second shot heard around

0:15

the world for the AI play, which is

0:19

basically the only thing that is

0:21

propping up GDP right now. It's not

0:25

great. I'm going to break down exactly

0:27

what's going on and why it's the second

0:29

dot heard around the world. Mind you,

0:32

this is coming from a company named

0:34

Oracle, which in case you don't recall,

0:38

Oracle is a person or thing regarded as

0:41

an infallible authority or guide on

0:44

something. Some would say like a seer or

0:46

like a, you know, a somebody who can

0:48

give divine or prophetic advice.

0:52

basically suggest a prophecy of what's

0:54

to come. You know, there's another

0:56

oracle known as the Oracle of Omaha who

0:59

also tells us that the stock market is

1:01

the most overvalued by 217% on the

1:04

Buffet indicator that it has ever been

1:06

in the history of market valuations.

1:08

Mind you, that snapshot is from June

1:11

30th. And let's just say we're up from

1:14

June 30th, so we're even more

1:16

overvalued.

1:18

All that notwithstanding, let's talk

1:20

about the two shots across the bow

1:25

and the shots heard around the world.

1:28

So, first you need to understand this

1:32

fancy piece of crap document called an

1:35

8K. It's another word for a current

1:38

report. Okay? It's basically a company

1:40

telling you, "Yo, we got some news.

1:42

We're just going to drop this here and

1:44

boom, it's been disclosed to you." So,

1:46

it's just a way for companies to cover

1:48

their legal basis for their disclosure

1:50

requirements. All right, fine. So,

1:53

what's in this disclosure that's so

1:55

fancy? Well, here you go. Okay. Material

1:59

service agreement. This bottom line,

2:03

we've talked about it before, so I'm

2:05

going to give you a very quick reminder.

2:06

This is a document that was filed about

2:08

30 days ago. H like 28

2:11

days ago. 28 days ago. Coreweave told us

2:15

that Cororeweave is forcing Nvidia to

2:18

buy unused

2:21

leases for Coreweave data centers that

2:25

are operated by or using Nvidia chips.

2:29

Now, the only way out of this is

2:31

bankruptcy. So in other words, Nvidia is

2:34

going to pay. Nvidia is going to pay to

2:36

buy back unused capacity at Core Reef

2:39

data centers all the way until 2032,

2:42

which is crazy. That's a long time. I

2:45

immediately made a video on this. I go,

2:46

"This is a shot to AI. This is bad." The

2:50

problem is a full copy of the material

2:53

services agreement will not be released

2:56

until after the next earnings report,

2:59

which we think will be November 12th.

3:02

though some have been saying the 14th.

3:04

So somewhere we'll we'll know as we get

3:06

closer. Somewhere between November 12th

3:07

and 14th. Mark your calendar for it. It

3:09

is a big deal. Okay? Because when this

3:14

came out, I made a video and said, "Hey,

3:16

this is a really big red flag because it

3:18

suggests that leasing out data centers

3:21

has no pricing power." And I made it

3:23

very clear that this is poopy dupy and

3:26

that we don't want poopy dupies. that we

3:28

want a frankly okay like as an owner of

3:32

Nvidia stock I just want Nvidia stock to

3:35

go up okay like there's that's that half

3:37

of me that's like please just make

3:39

Nvidia stock go up okay I like AI boom

3:42

the AI boom's great it's supporting our

3:44

GDP is fantastic but that's not the

3:48

analyst side the analyst side says this

3:50

is a big red flag to PP it means that

3:53

renting out data centers don't have

3:56

pricing hour. Now, what did we learn

3:59

today and why did the market tank today?

4:02

Well, the market tank today because

4:04

Oracle, who's a heavily indebted

4:06

company, just came out and gave us some

4:09

numbers that suck.

4:12

So, over the last 3 months, Oracle has

4:15

generated $900 million in rentals of

4:18

servers. Their executives told us that

4:22

in the future them renting servers to

4:24

OpenAI would generate $381 billion of

4:28

revenues.

4:32

That's

4:34

105

4:35

times how much money they're making

4:37

currently from the rentals of servers.

4:40

So the executives are like, "Yeah, we're

4:42

making $900 right now. Uh that's

4:44

actually going to be uh 381,

4:47

you know, thousands uh or we're making

4:50

$900 per quarter. So that's like 300

4:54

$3,600 per uh year. Uh yeah, that's

4:58

actually going to 100x in the future. Uh

5:01

trust us, bro.

5:03

Okay. Well, great. So then of course

5:06

Oracle skyrockets on the news, but now

5:09

we're getting internal leaked

5:10

information that while Oracle on that

5:14

$900 million sees a $125 million like

5:18

gross profit, which is about a 14% gross

5:21

margin,

5:22

Oracle actually lost $und00 million

5:26

renting out Blackwell chips, which are

5:28

supposed to be the latest and greatest

5:30

chips from Nvidia, which supposedly have

5:32

all of this insane demand. Now maybe for

5:36

a cashrich company that's not that big

5:38

of a deal. It would just be considered

5:39

the cost of doing business. But then

5:41

when we actually look at the balance

5:44

sheet for Oracle, we start getting a

5:45

little bit concerned. And I've warned

5:47

about this on the channel before that

5:49

Oracle is one of the highly indebted

5:51

ones. And I've actually applauded Amazon

5:54

for not being as highly indebted as

5:56

Oracle. And this is, you know, what's

5:58

leading the market to tank today, right?

6:00

Uh the problem is Amazon

6:03

literally today said they are opening

6:07

new lines of credit with banks to

6:11

finance more AI investment. I'm like ah

6:15

not them as well. Okay. Anyway, so

6:18

everybody's taking on debt, but I want

6:20

you to look at this balance sheet and

6:21

tell me what this balance sheet looks

6:22

like to you. Tell me if this is good.

6:24

They have $19 billion to pay bills. So,

6:28

imagine you had $19 in your bank account

6:31

and then you had $27 of bills to pay

6:34

this year. Well, you know, in the next

6:37

12 months, that means you're upside down

6:39

already. Okay? They've got $27 billion

6:40

in bills to pay. They've got $19 billion

6:44

in cash and receivables, assuming they

6:46

actually get all these receivables. Keep

6:48

in mind, in their current liabilities,

6:49

I'm excluding deferred revenues that

6:50

they've already counted. Now, on top of

6:53

that, they have about $100 billion of

6:56

debt. not including their leases.

7:00

So, they're already heavily in debt. On

7:03

top of that, when you look at their cash

7:04

flow statement, yes, they currently have

7:08

cash flow from operating activities, but

7:11

they're spending more than all of it. If

7:14

you look at their free cash flow by

7:16

taking this number and then subtracting

7:18

their capex, they're actually negative

7:20

$400 million in 3 months on free cash

7:25

flow. So, we're negative free cash flow.

7:27

How are we coming up with that money?

7:29

We're taking on $800 million more

7:32

financing in just 3 months. We're

7:34

issuing $1.1 billion of stock. So,

7:36

they're diluting their investors.

7:39

Uh, and they're using some of it to pay

7:42

dividends and the rest to finance their

7:44

negative free cash flow because they're

7:45

so in debt. So, this is not like a very

7:48

financially sound company even though

7:50

it's massive, which is kind of scary.

7:52

It's almost gets to like too big to fail

7:54

level. So then you look at their

7:55

margins. Their margins have collapsed

7:58

240 basis points from last year. Last

8:01

year they net about $2.9 billion, which

8:04

what you'll notice is exactly the same

8:07

as what they net this year. See, 2.9

8:09

billion, 2.9 billion, right? But wait a

8:12

minute, what was the revenue? Oh my

8:13

gosh, last year was 13.3 billion. Now

8:16

it's 14.9

8:18

and they net the same. So their revenue

8:20

blew up, but they net the same. That's

8:23

because margin's collapsing. Okay, why

8:26

is margin collapsing? They can't make

8:28

money on these damn data centers, which

8:30

is bad because remember the circular

8:33

arrangement we talked about yesterday?

8:36

Let me show you a picture of

8:40

the circular

8:42

scam.

8:44

Oh, sorry. I actually meant to pull up

8:48

my iPad a bit to remind you all of the

8:51

circular nature of the relationship that

8:54

we talked about yesterday which is only

8:56

amplified mind you by the amount of

8:59

deals that open AAI is doing. Look at

9:02

this chart from the information just to

9:05

understand how critical open AAI here

9:08

is. Open AAI has a deal with SoftBank

9:11

has a deal with Oracle who have deals

9:13

with Nvidia and OpenAI also has a deal

9:15

with Nvidia. Nvidia has deals with

9:17

Cororeweave. Cororeef has deals with

9:19

Nvidia, mind you, to lease back unused

9:22

chip capacity. You've got deals with

9:24

Microsoft who has deals with Coree. And

9:26

OpenAI also has deals with Microsoft.

9:28

OpenAI has chip deals with Broadcom to

9:30

make a custom chip. OpenAI has deals

9:32

with Google. Uh oh, Google has deals

9:36

with Anthrope. Google has deals with

9:38

Meta, Amazon with Anthrop. And of

9:40

course, OpenAI with Amazon. Uh, and

9:46

I don't know what's missing here.

9:48

Thanks, Financial Times. Sandbagging me

9:50

here. But anyway, the point is

9:54

all of this is fuel to the fire that

9:58

Oracle is now losing money on these

10:00

rentals when we look at this, which is

10:03

like this disastrous circle. But

10:05

understand, I'm going to simplify this,

10:07

okay? Nvidia invests a billion dollars

10:10

for every gigawatt deployed into OpenAI.

10:13

So OpenAI gets money when they deploy a

10:15

gigawatt. Okay. In the second half of

10:18

next year, OpenAI is going to have their

10:21

first gigawatt deployed, which is when

10:23

Nvidia is going to give OpenAI $10

10:25

billion. So in the second half of next

10:27

year, Nvidia gives OpenAI $10 billion.

10:31

After they get that $10 billion, in the

10:33

second half, AMD says they're going to

10:35

begin their data center deployments with

10:38

the money that OpenAI

10:40

has to spend on AMD chips.

10:44

And then of course AMD will get revenues

10:47

which will lead AMD stock to go up which

10:50

open door or sorry open AAI owns shares

10:53

of warrants of

10:56

and Nvidia benefits from the value of

10:59

OpenAI going up because it gives OpenAI

11:01

more capital to spend money on Nvidia

11:03

chips as well. It's insane. Okay, that's

11:06

just a reminder from yesterday. Add that

11:08

to the news that we got today from the

11:10

information. It's not good. So

11:13

understand that Oracle is now seeing

11:16

losses of a hund00 million from renting

11:18

Blackwell chips. That's crazy. There's

11:21

also no standard on the depreciation for

11:24

these products, which is also not good.

11:28

Understand that depreciation, and I've

11:31

made videos on this, is fantastic to add

11:34

back in when prices are going up, but

11:38

these are depreciating assets. chips are

11:41

going to go down in value over time.

11:44

Over the long term, a chip goes to zero.

11:48

Like I I just want to make this like

11:50

critically clear, okay? Anything like a

11:54

chip, a car, right? These things go to

11:57

zero over time. So, uh let's write that

12:00

down here. Let's go. Um, let's write

12:02

down ships,

12:05

AI, or I shouldn't say AI, uh, cars,

12:10

planes,

12:12

boats,

12:14

butter,

12:16

you know, for those of you who are

12:17

course members and understand what that

12:18

is. All of this trends down

12:22

to zero over time. It goes it always

12:26

goes to zero. That's what history tells

12:28

us. Okay? There's there's no there's no

12:30

debating that these things go to zero

12:33

and so you depreciate them to take them

12:35

to zero. Of course, then you know people

12:37

add back in depreciation but they're

12:38

like because they're like oh value. The

12:41

only place it makes sense to add back in

12:43

depreciation

12:45

is real estate

12:48

because houses over time are an

12:50

inflation hedge that go up in value. All

12:54

right? They hedge inflation. they remove

12:59

you from the dollar, right? So, housing,

13:03

it makes sense to add back depreciation

13:06

because it goes up over the long term.

13:08

Everybody over here adding back in

13:10

depreciation for a depreciating asset is

13:13

losing their mind and there is no

13:16

standard definition of what's going on

13:19

with depreciation at these companies.

13:21

Now, why is that a problem? Well,

13:23

because Oracle themselves uh via this

13:26

leaked information on depreciation, let

13:28

me show you. Take a look at this. Here

13:30

we go. 14% gross margin, which is them

13:34

just taking the 120 million of gross

13:36

profit divided by 900 figures about a

13:38

14% gross margin, which mind you is is

13:41

lower than Oracle's net right now,

13:44

right? Oracle's net margins. The more AI

13:47

data centers they do, the more their net

13:49

margins are going down. Their net margin

13:51

right now is 19.6%. Right? Their gross

13:55

margin on AI chips is 14%. But that

14:00

doesn't include potentially another 7%

14:04

of depreciation that they could include

14:07

and but there's no standard definition

14:09

on how people calculate this. So it's

14:12

crazy. These are apparently internal

14:14

documents uh from Oracle. And so it's

14:17

all a fugazi Like let me let

14:21

me sorry I shouldn't say bad ones but

14:25

and and mind you like I'm losing money

14:29

right now. Okay. Like this is this sucks

14:31

dude. What is like you see this? You see

14:32

Nvidia down 50 basis points. Bro, that's

14:35

costing me money by the second film.

14:38

Like honestly we were filming like

14:40

looking at this information like right

14:42

here on the live stream and I'm like oh

14:44

this is really bad for Nvidia. Like I

14:46

could have sold right here. I just lost

14:49

another three bucks per share at, you

14:51

know, over 13,000 shares or whatever I

14:53

got. So like, damn me. But I must make

14:57

this video.

14:59

It's okay. It's okay. So like I I want

15:02

to shill AI.

15:05

But my nature is I got I got to tell the

15:08

world what I'm seeing. And when I see

15:09

bull crap, I'm going to talk about the

15:11

bull crap. And so here's the thing,

15:13

okay? You have to understand that these

15:16

companies earnings are fugazed. Okay,

15:19

here's how. I'm going to show you the

15:22

fugaze how it works. Okay, watch this.

15:24

Watch this. So, uh I'm going to go

15:27

pretend I'm Microsoft, let's say. Okay,

15:31

this is

15:32

uh fake. Don't sue me, bro. Example

15:36

given. Okay, do people understand what

15:38

EG is? It stands for example given.

15:41

Okay, if you just learned something, hit

15:44

subscribe. Oh my gosh, dude. I need I

15:47

need Alamo coffee for this. Remember the

15:50

Alamo.

15:54

All right, this is important.

15:57

So, fake numbers. Don't sue me, bro

16:00

Microsoft. So, hey guys. Uh we just

16:05

spent

16:07

uh10

16:09

billion dollar

16:11

on Invenia chips

16:15

and our data center build out. Yeah.

16:18

Yeah.

16:19

>> Oh, guys, guys, guys. Uh we just made

16:24

um

16:26

I don't know, let's go with uh we just

16:30

made,

16:31

how do we want to put it? Uh, let's just

16:33

say, I don't know, let's just say they

16:35

made $2 billion. We just made $2 billion

16:38

in revenue, you know, recurring revenue

16:41

or what, right? Well, technically,

16:45

you know, you and I on a cash basis,

16:48

we're like, "All right, so you just

16:49

spent $10 billion on chips to make 2

16:51

billion. You just lostg

16:54

billion, right?" But that's not how it

16:57

works because these companies say,

17:00

"Well, these $10 billion of chips, we're

17:03

actually going to depreciate them over 7

17:06

years. So, we're only going to expense

17:09

these to the tune of $1.4 billion."

17:16

So, actually, guys, we didn't just lose

17:19

$8 billion on a depreciating asset that

17:22

Oracle can't even rent out for a profit.

17:25

We actually

17:27

made $540 million

17:31

cuz we're top G's, bro.

17:35

So, so you literally have these

17:37

companies today. Literally Micros,

17:39

they're all doing it. Microsoft, Meta,

17:42

Amazon, dude, our entire GDP

17:45

is is like based on these knuckleheads

17:48

right now. And these are like toothpick

17:49

foundations, okay? They're going, "Guys,

17:52

we just spent all this money on stuff

17:55

that's going to lose value if God g. But

17:58

don't worry, my uh accounting."

18:11

So, like I said, like when I see stuff

18:15

like this to win big deals with AI

18:17

customers, Oracle and other cloud

18:19

providers were offering heavy discounts

18:21

on GPU prices and Azure is seeing uh

18:26

margin pressures at Microsoft because

18:28

companies have to price so aggressively.

18:30

And on top of that, when I see

18:32

utilization is trash. Where was it?

18:35

Utilization. Utiliz here. utilization of

18:38

Oracle's GPU servers range between 60

18:40

and 90%.

18:42

That's bad.

18:44

That's bad. Like, understand this, okay?

18:48

Uh, house attack is at 100% occupancy.

18:52

Our tenants pay rent and I get pissed

18:55

when properties are vacant. Obviously,

18:57

when we're renovating properties,

18:59

they're vacant and the goal is rent them

19:01

out as soon as possible. Now, most

19:04

apartment buildings operate at like 92

19:07

to 94% occupancy. So, you have a vacancy

19:10

factor you included. Okay, that's like

19:14

apartment building world 92 94%. You

19:17

know, we're at 100% on our apartment

19:18

buildings. But

19:20

the utilization for these chips which

19:22

can run 247 and aren't reliant on people

19:27

is between 60 and 90%. So take the

19:30

midpoint 75%.

19:34

That's horrible. That means 25% of the

19:37

time all these data centers that you

19:39

currently have on average are doing

19:42

nothing.

19:44

Which is really interesting because if

19:46

25% of the time the data centers you

19:48

currently have are doing nothing, why

19:50

are we building so many more?

19:54

And this is where the problem comes.

19:57

Because if all of a sudden we're like,

20:00

you know what? We don't need to build

20:01

anymore. Here's how it all collapses.

20:04

It's like we're writing the 2026

20:06

textbook on the collapse. Here's how it

20:08

works. Nvidia invests $10 billion into

20:11

OpenAI for every gigawatt deployed. Oh,

20:15

but wait a minute. What happens for a

20:18

moment? What happens if we don't get

20:22

that gigawatt deployed? Oh no, no

20:26

deployment. Oh crap.

20:29

Okay. Well, wait a minute. Then that

20:33

means Open AI won't get their $10

20:36

billion from Nvidia. Gone. Which, wait a

20:40

minute. That means Open AI doesn't have

20:43

the money to invest in AMD to begin the

20:46

AMD buildout. Oh crap. Which means AMD

20:51

doesn't get the money. Which means

20:54

Nvidia doesn't get the revenue.

20:57

Which means Nvidia,

21:00

oh crap, can't invest anymore,

21:05

which means that OpenAI stock or that

21:07

AMD stock isn't going to go up, which

21:10

means OpenAI doesn't have more money,

21:12

which means no more more Nvidia.

21:15

Oh crap. I'm left with a shell. Why not

21:18

advertise these things that you told us

21:20

here? I feel like nobody else knows

21:21

about this. We'll we'll try a little

21:23

advertising and see how it goes.

21:24

>> Congratulations, man. You have done so

21:26

much. People love you. People look up to

21:27

you.

21:28

>> Kevin Praath there, financial analyst

21:30

and YouTuber. Meet Kevin. Always great

21:32

to get your take.

UNLOCK MORE

Sign up free to access premium features

INTERACTIVE VIEWER

Watch the video with synced subtitles, adjustable overlay, and full playback control.

SIGN UP FREE TO UNLOCK

AI SUMMARY

Get an instant AI-generated summary of the video content, key points, and takeaways.

SIGN UP FREE TO UNLOCK

TRANSLATE

Translate the transcript to 100+ languages with one click. Download in any format.

SIGN UP FREE TO UNLOCK

MIND MAP

Visualize the transcript as an interactive mind map. Understand structure at a glance.

SIGN UP FREE TO UNLOCK

CHAT WITH TRANSCRIPT

Ask questions about the video content. Get answers powered by AI directly from the transcript.

SIGN UP FREE TO UNLOCK

GET MORE FROM YOUR TRANSCRIPTS

Sign up for free and unlock interactive viewer, AI summaries, translations, mind maps, and more. No credit card required.